DD1_MCS_report_hapto_8_27_19
Dopfer
8/27/2019
Set working directory
Load libraries
load .RData
## Observations: 712
## Variables: 72
## $ CowID <fct> 276000941090648, 276000951311013, 2760009…
## $ Cow.Birth.Date <date> 2007-10-19, 2016-05-31, 2016-05-31, 2012…
## $ Farm.Name <fct> Stadler, Stadler, Stadler, Stadler, Stadl…
## $ Farm.No <fct> 14184137142, 14184137142, 14184137142, 14…
## $ Cow.Breed <fct> 01, 01, 01, 01, 01, 01, 01, 01, 01, 01, 0…
## $ Cow.Exit.Date <date> NA, NA, NA, NA, NA, NA, NA, NA, 2018-09-…
## $ BS.DATE <date> 2018-11-06, 2018-07-24, 2018-08-07, 2018…
## $ BS.Month <fct> 11-2018, 07-2018, 08-2018, 12-2018, 12-20…
## $ BS.MS.Date.Difference <dbl> 1, 1, 0, 1, 1, 1, 0, 1, 1, 1, 1, 1, 0, 1,…
## $ BS.DIM <dbl> 21, 15, 29, 16, 16, 28, 11, 22, 6, 13, 13…
## $ Hapto.Result <dbl> 597.5, 280.0, 741.5, 280.0, 1903.5, 11584…
## $ Hapto.Log <dbl> 6.392754, 5.634790, 6.608675, 5.634790, 7…
## $ Hapto.0.1 <fct> 0, 0, 0, 0, 1, 1, 1, 0, 1, 1, 0, 1, 1, 0,…
## $ Hapto0.35 <fct> 0, 0, 0, 0, 0, 1, 1, 0, 1, 1, 0, 1, 1, 0,…
## $ Hapto.HML <fct> 0, 0, 0, 0, 0, 1, 0, 0, 1, 1, 0, 0, 2, 0,…
## $ Hapto.HL <fct> 1, 1, 1, 1, 1, NA, 1, 1, NA, NA, 1, 1, 3,…
## $ BS.NEFA <dbl> 0.23, 0.68, 0.48, 0.32, 0.33, 0.22, 0.50,…
## $ BS.BHBA <dbl> 1.35, 1.49, 1.78, 1.46, 1.26, 0.98, 0.71,…
## $ MS.Date.Taken <date> 2018-11-05, 2018-07-23, 2018-08-07, 2018…
## $ MS.Time.Taken <chr> "06:42:00", "12:00:00", "12:00:00", "01:0…
## $ Calving.Date <date> 2018-10-16, 2018-07-09, 2018-07-09, 2018…
## $ Calving.No <dbl> 9, 1, 1, 5, 7, 9, 4, 6, 2, 5, 1, 1, 1, 6,…
## $ Hfr.or.Cow <fct> cow, hfr, hfr, cow, cow, cow, cow, cow, c…
## $ MS.DIM <dbl> 20, 14, 29, 15, 15, 27, 11, 21, 5, 12, 12…
## $ MS.Date.Analyzed <date> 2018-11-07, 2018-07-26, 2018-08-08, 2018…
## $ Milk.Yield <dbl> 39.9, 23.3, NA, 41.6, 30.4, 41.6, NA, 40.…
## $ MS.Milk.Yield <dbl> 21.2, 23.3, 14.9, 20.7, 11.9, 17.0, 12.5,…
## $ MS.Fat <dbl> 4.32, 4.70, 4.21, 6.12, 5.04, 5.61, 3.82,…
## $ MS.Protein <dbl> 2.79, 3.05, 2.85, 3.42, 2.89, 2.76, 3.30,…
## $ MS.S.Cell.Count <dbl> 16, 298, 231, 21, 10, 21, 19, 10, 249, 66…
## $ MS.Lactose <dbl> 4.73, 4.86, 4.97, 4.82, 4.81, 4.80, 4.83,…
## $ MS.Urea <dbl> 24.4, 52.2, 26.0, 28.1, 20.0, 30.0, 16.5,…
## $ MS.pH <dbl> 6.64, 6.65, 6.68, 6.66, 6.66, 6.64, 6.70,…
## $ MS.Acetone <dbl> 10, 140, 140, 170, 120, 60, 90, NA, 270, …
## $ MS.BHBA <dbl> 20, 90, 90, 170, 90, 70, 50, 10, 170, 40,…
## $ MS.NEFA <dbl> 6.30, 9.90, 8.00, 4.90, 4.00, 5.20, 6.60,…
## $ MS.NSFA <dbl> 14.2, 19.6, 15.9, 22.9, 19.9, 20.1, 15.7,…
## $ MS.MFA <dbl> 1.142, 1.653, 1.319, 1.878, 1.657, 1.641,…
## $ MS.PFA <dbl> 0.144, 0.194, 0.137, 0.207, 0.182, 0.193,…
## $ MS.SFA <dbl> 2.653, 2.569, 2.435, 3.726, 2.868, 3.375,…
## $ MS.Palmeic <dbl> 1.172, 1.094, 1.115, 1.422, 1.235, 1.410,…
## $ MS.Stearine <dbl> 0.471, 0.653, 0.534, 0.735, 0.637, 0.636,…
## $ MS.Oleic <dbl> 1.061, 1.596, 1.199, 1.802, 1.534, 1.582,…
## $ CE.Date <date> 2018-11-06, 2018-07-24, 2018-08-07, 2018…
## $ CE.Temp <dbl> 38.2, 38.7, 38.4, 38.2, 38.5, 38.0, 38.1,…
## $ CE.Environ.Temp <dbl> 8.0, 19.3, 22.4, 3.9, 11.9, 6.0, 22.4, 3.…
## $ CE.Fat.Level <dbl> 12, 19, 12, 16, 13, 12, 19, 20, 17, 35, 2…
## $ CE.Skin.Dehydration <fct> phy, phy, phy, phy, phy, phy, phy, phy, p…
## $ CE.Stom.Tension <fct> phy, phy, phy, phy, phy, phy, phy, phy, p…
## $ CE.Stom.Layering <fct> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,…
## $ CE.Stom.Fluid.Movement <fct> obb, obb, obb, obb, obb, obb, obb, obb, o…
## $ CE.Stom.Ping <fct> obb, obb, obb, obb, obb, obb, obb, obb, o…
## $ CE.Stom.Fullness <fct> norm, abnorm, abnorm, abnorm, abnorm, nor…
## $ CE.Stom.Noise.Freq <fct> NA, NA, NA, NA, NA, NA, NA, NA, bis_2, NA…
## $ CE.Waste.Consistency <fct> norm, norm, norm, norm, norm, norm, norm,…
## $ CE.Waste.Digestion <fct> maess, gut, gut, maess, gut, maess, gut, …
## $ CE.Locomotion.Score <fct> 3, 1, 1, 2, 1, 3, 2, 2, 3, 2, 1, 1, 2, 1,…
## $ CE.Lame <fct> 1, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0,…
## $ BS.NEFA.Log <dbl> -1.4696760, -0.3856625, -0.7339692, -1.13…
## $ BS.BHBA.Log <dbl> 0.30010459, 0.39877612, 0.57661336, 0.378…
## $ MS.S.Cell.Count.Log <dbl> 2.772589, 5.697093, 5.442418, 3.044522, 2…
## $ MS.BHBA.Log <dbl> 2.995732, 4.499810, 4.499810, 5.135798, 4…
## $ MS.Acetone.Log <dbl> 2.302585, 4.941642, 4.941642, 5.135798, 4…
## $ BS.NEFA.0.7 <fct> 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0,…
## $ BS.BHBA.1.2 <fct> 1, 1, 1, 1, 1, 0, 0, 1, 1, 0, 1, 0, 0, 0,…
## $ Hapto0.35.2 <fct> 0, 0, 0, 0, 0, 1, 1, 0, 1, 1, 0, 1, 1, 0,…
## $ BS.Month2 <fct> 11-2018, 07-2018, 08-2018, 12-2018, 12-20…
## $ BS.Month_warm <fct> 0, 1, 1, 0, 0, 0, 1, 0, 1, 1, 0, 1, 1, 0,…
## $ CE.Stom.Fullness01 <fct> 0, 1, 1, 1, 1, 0, 0, 0, 0, 1, 1, 1, 0, 0,…
## $ CE.Stom.Tension01 <fct> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
## $ clusterHapto5 <int> 1, 2, 2, 4, 2, 1, 3, 4, 1, 3, 3, 3, 2, 4,…
## $ clusterHapto <int> 2, 2, 2, 3, 2, 1, 2, 3, 1, 2, 2, 2, 2, 3,…Explore Month of sampling and warm (months 7, 8, 9) vs cooler months (months: 10, 11, 12, and 6) for effect on Hapto.Log
##
## Call:
## lm(formula = Hapto.Log ~ BS.Month, data = hp9)
##
## Residuals:
## Min 1Q Median 3Q Max
## -3.153 -1.604 -1.002 1.010 7.468
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 6.3227 0.3127 20.219 < 2e-16 ***
## BS.Month07-2018 1.3364 0.4050 3.300 0.00102 **
## BS.Month08-2018 1.8506 0.3805 4.863 1.43e-06 ***
## BS.Month09-2018 1.6428 0.4106 4.001 6.97e-05 ***
## BS.Month10-2018 0.8946 0.3778 2.368 0.01816 *
## BS.Month11-2018 0.6158 0.3787 1.626 0.10439
## BS.Month12-2018 0.4510 0.4372 1.032 0.30264
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 2.482 on 705 degrees of freedom
## Multiple R-squared: 0.0532, Adjusted R-squared: 0.04515
## F-statistic: 6.603 on 6 and 705 DF, p-value: 8.371e-07
##
## Call:
## lm(formula = Hapto.Log ~ BS.Month_warm, data = hp9)
##
## Residuals:
## Min 1Q Median 3Q Max
## -3.0198 -1.5523 -1.1927 0.8327 7.4966
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 6.9099 0.1242 55.624 < 2e-16 ***
## BS.Month_warm1 1.0515 0.1880 5.594 3.16e-08 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 2.488 on 710 degrees of freedom
## Multiple R-squared: 0.04222, Adjusted R-squared: 0.04087
## F-statistic: 31.3 on 1 and 710 DF, p-value: 3.162e-08
Start: baseline model and full model for linear regression 
## Analysis of Deviance Table (Type III Wald chisquare tests)
##
## Response: Hapto.Log
## Chisq Df Pr(>Chisq)
## (Intercept) 1.7316 1 0.1882102
## MS.Milk.Yield 1.9544 1 0.1621112
## MS.NEFA 0.4721 1 0.4920337
## MS.DIM 0.5416 1 0.4617607
## MS.Fat 1.4057 1 0.2357791
## MS.Protein 10.5874 1 0.0011386 **
## MS.Lactose 2.9903 1 0.0837655 .
## MS.Acetone.Log 0.3383 1 0.5608100
## MS.Urea 2.8482 1 0.0914773 .
## MS.S.Cell.Count.Log 16.0056 1 6.316e-05 ***
## MS.pH 0.0006 1 0.9798728
## CE.Environ.Temp 0.0992 1 0.7527752
## Hfr.or.Cow 3.1194 1 0.0773656 .
## BS.Month_warm 2.2656 1 0.1322752
## CE.Skin.Dehydration 0.3119 1 0.5765167
## CE.Stom.Tension01 1.3232 1 0.2500136
## CE.Stom.Layering 0.1856 1 0.6665866
## CE.Stom.Fullness01 0.3702 1 0.5429081
## CE.Stom.Noise.Freq 2.0099 2 0.3660696
## CE.Waste.Consistency 1.4452 3 0.6949762
## CE.Lame 11.6515 1 0.0006415 ***
## BS.NEFA.0.7 0.4927 1 0.4827087
## BS.BHBA 2.3703 1 0.1236604
## Cow.Breed 0.5619 1 0.4535058
## BS.NEFA.0.7:BS.BHBA 0.0239 1 0.8770401
## MS.Milk.Yield:BS.BHBA 0.3536 1 0.5520542
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Note: there are still those linear shapes in the residual plots indicating that we are missing information explaning the hapto outcomes
Final model from Backward step elimination
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula:
## Hapto.Log ~ MS.Milk.Yield + MS.Protein + MS.Lactose + MS.S.Cell.Count.Log +
## Hfr.or.Cow + BS.Month_warm + CE.Stom.Tension01 + CE.Stom.Fullness01 +
## CE.Lame + BS.NEFA.0.7 + (1 | CowID:Farm.No)
## Data: hp9
##
## REML criterion at convergence: 3133.7
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.7468 -0.5993 -0.2304 0.4390 3.2446
##
## Random effects:
## Groups Name Variance Std.Dev.
## CowID:Farm.No (Intercept) 0.6406 0.8004
## Residual 4.1093 2.0271
## Number of obs: 712, groups: CowID:Farm.No, 423
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 12.16543 2.90897 579.30000 4.182 3.34e-05 ***
## MS.Milk.Yield -0.03940 0.01437 655.30000 -2.741 0.00629 **
## MS.Protein -0.62638 0.28879 638.50000 -2.169 0.03045 *
## MS.Lactose -1.01367 0.49800 576.20000 -2.035 0.04226 *
## MS.S.Cell.Count.Log 0.48289 0.06990 561.00000 6.908 1.33e-11 ***
## Hfr.or.Cowhfr 0.81732 0.20284 396.10000 4.029 6.71e-05 ***
## BS.Month_warm1 0.93144 0.17766 469.40000 5.243 2.40e-07 ***
## CE.Stom.Tension011 0.99395 0.32674 696.70000 3.042 0.00244 **
## CE.Stom.Fullness011 -0.58044 0.17424 699.90000 -3.331 0.00091 ***
## CE.Lame1 1.23968 0.26958 646.90000 4.599 5.12e-06 ***
## BS.NEFA.0.71 1.02456 0.22097 663.70000 4.637 4.27e-06 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) MS.M.Y MS.Prt MS.Lct MS.S.C Hfr..C BS.M_1 CE.S.T CE.S.F
## MS.Milk.Yld -0.127
## MS.Protein -0.547 0.134
## MS.Lactose -0.935 -0.001 0.246
## MS.S.Cl.C.L -0.392 0.037 0.003 0.353
## Hfr.r.Cwhfr 0.032 0.218 0.145 -0.125 -0.081
## BS.Mnth_wr1 0.033 -0.081 0.092 -0.088 -0.084 0.000
## CE.Stm.T011 0.013 0.015 -0.049 -0.007 -0.016 -0.001 0.102
## CE.Stm.F011 0.061 0.065 -0.042 -0.106 0.012 -0.125 -0.028 -0.009
## CE.Lame1 -0.083 0.037 0.120 0.036 -0.007 0.088 -0.067 -0.137 0.088
## BS.NEFA.0.7 -0.142 0.088 0.157 0.080 0.011 -0.047 -0.070 0.042 0.193
## CE.Lm1
## MS.Milk.Yld
## MS.Protein
## MS.Lactose
## MS.S.Cl.C.L
## Hfr.r.Cwhfr
## BS.Mnth_wr1
## CE.Stm.T011
## CE.Stm.F011
## CE.Lame1
## BS.NEFA.0.7 -0.066
Repeat multiple variable analysis as logistic regression with Hapto0.35 as outcome variable
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula:
## Hapto0.35 ~ MS.Milk.Yield + MS.Protein + MS.Lactose + MS.S.Cell.Count.Log +
## Hfr.or.Cow + BS.Month_warm + CE.Stom.Tension01 + CE.Stom.Fullness01 +
## CE.Lame + BS.NEFA.0.7 + BS.BHBA + BS.NEFA.0.7 * BS.BHBA +
## MS.Milk.Yield * BS.BHBA + Cow.Breed + (1 | CowID:Farm.No)
## Data: hp9
##
## AIC BIC logLik deviance df.resid
## 686.9 760.0 -327.5 654.9 696
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.1861 -0.5330 -0.3407 0.3916 6.8042
##
## Random effects:
## Groups Name Variance Std.Dev.
## CowID:Farm.No (Intercept) 5.962e-15 7.721e-08
## Number of obs: 712, groups: CowID:Farm.No, 423
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 6.08996 3.45404 1.763 0.077877 .
## MS.Milk.Yield -0.05925 0.03007 -1.971 0.048770 *
## MS.Protein -0.77501 0.35488 -2.184 0.028972 *
## MS.Lactose -1.27327 0.57467 -2.216 0.026715 *
## MS.S.Cell.Count.Log 0.42384 0.07881 5.378 7.54e-08 ***
## Hfr.or.Cowhfr 0.80479 0.22288 3.611 0.000305 ***
## BS.Month_warm1 0.83619 0.20152 4.149 3.33e-05 ***
## CE.Stom.Tension011 0.52014 0.37065 1.403 0.160516
## CE.Stom.Fullness011 -0.49354 0.20765 -2.377 0.017467 *
## CE.Lame1 1.25210 0.28322 4.421 9.83e-06 ***
## BS.NEFA.0.71 0.77156 0.42691 1.807 0.070715 .
## BS.BHBA -0.81052 0.49988 -1.621 0.104921
## Cow.Breed02 0.03740 0.22505 0.166 0.868011
## BS.NEFA.0.71:BS.BHBA 0.15969 0.36375 0.439 0.660650
## MS.Milk.Yield:BS.BHBA 0.02970 0.02362 1.258 0.208537
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1##
## Correlation matrix not shown by default, as p = 15 > 12.
## Use print(x, correlation=TRUE) or
## vcov(x) if you need it
## Analysis of Deviance Table (Type III Wald chisquare tests)
##
## Response: Hapto0.35
## Chisq Df Pr(>Chisq)
## (Intercept) 3.1087 1 0.0778769 .
## MS.Milk.Yield 3.8832 1 0.0487702 *
## MS.Protein 4.7692 1 0.0289725 *
## MS.Lactose 4.9091 1 0.0267151 *
## MS.S.Cell.Count.Log 28.9220 1 7.535e-08 ***
## Hfr.or.Cow 13.0386 1 0.0003051 ***
## BS.Month_warm 17.2175 1 3.333e-05 ***
## CE.Stom.Tension01 1.9694 1 0.1605159
## CE.Stom.Fullness01 5.6488 1 0.0174669 *
## CE.Lame 19.5448 1 9.827e-06 ***
## BS.NEFA.0.7 3.2664 1 0.0707146 .
## BS.BHBA 2.6291 1 0.1049209
## Cow.Breed 0.0276 1 0.8680106
## BS.NEFA.0.7:BS.BHBA 0.1927 1 0.6606496
## MS.Milk.Yield:BS.BHBA 1.5816 1 0.2085366
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula:
## Hapto0.35 ~ MS.Milk.Yield + MS.Protein + MS.Lactose + MS.S.Cell.Count.Log +
## Hfr.or.Cow + BS.Month_warm + CE.Stom.Fullness01 + CE.Lame +
## BS.NEFA.0.7 + BS.BHBA.Log + (1 | CowID:Farm.No)
## Data: hp9
##
## AIC BIC logLik deviance df.resid
## 679.8 734.6 -327.9 655.8 700
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.2034 -0.5258 -0.3342 0.3889 5.6275
##
## Random effects:
## Groups Name Variance Std.Dev.
## CowID:Farm.No (Intercept) 0 0
## Number of obs: 712, groups: CowID:Farm.No, 423
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 5.75753 3.35422 1.717 0.086070 .
## MS.Milk.Yield -0.02844 0.01778 -1.600 0.109666
## MS.Protein -0.81294 0.35211 -2.309 0.020958 *
## MS.Lactose -1.35465 0.56911 -2.380 0.017299 *
## MS.S.Cell.Count.Log 0.42946 0.07857 5.466 4.60e-08 ***
## Hfr.or.Cowhfr 0.79757 0.22130 3.604 0.000313 ***
## BS.Month_warm1 0.78677 0.19851 3.963 7.39e-05 ***
## CE.Stom.Fullness011 -0.48396 0.20598 -2.350 0.018797 *
## CE.Lame1 1.28389 0.27720 4.632 3.63e-06 ***
## BS.NEFA.0.71 0.91805 0.23945 3.834 0.000126 ***
## BS.BHBA.Log -0.50264 0.21676 -2.319 0.020401 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) MS.M.Y MS.Prt MS.Lct MS.S.C Hfr..C BS.M_1 CE.S.F CE.Lm1
## MS.Milk.Yld -0.134
## MS.Protein -0.576 0.140
## MS.Lactose -0.932 -0.002 0.277
## MS.S.Cl.C.L -0.357 0.035 -0.048 0.325
## Hfr.r.Cwhfr 0.034 0.222 0.130 -0.142 -0.001
## BS.Mnth_wr1 0.022 -0.084 0.083 -0.089 -0.013 0.044
## CE.Stm.F011 0.075 0.064 -0.023 -0.120 -0.045 -0.166 -0.051
## CE.Lame1 -0.080 0.021 0.100 0.020 0.091 0.159 0.022 0.072
## BS.NEFA.0.7 -0.063 0.091 0.077 -0.004 0.063 -0.021 -0.029 0.181 -0.032
## BS.BHBA.Log -0.240 -0.043 0.276 0.210 -0.014 0.021 -0.004 -0.014 0.053
## BS.NEF
## MS.Milk.Yld
## MS.Protein
## MS.Lactose
## MS.S.Cl.C.L
## Hfr.r.Cwhfr
## BS.Mnth_wr1
## CE.Stm.F011
## CE.Lame1
## BS.NEFA.0.7
## BS.BHBA.Log -0.278
Start k-means clustering analysis collect all ssign variables into hp10 and format factor variables as numeric dummy variables; check pair-wise correlations again
## Observations: 712
## Variables: 10
## $ MS.Milk.Yield <dbl> 21.2, 23.3, 14.9, 20.7, 11.9, 17.0, 12.5, 11…
## $ MS.Protein <dbl> 2.79, 3.05, 2.85, 3.42, 2.89, 2.76, 3.30, 3.…
## $ MS.Lactose <dbl> 4.73, 4.86, 4.97, 4.82, 4.81, 4.80, 4.83, 4.…
## $ MS.S.Cell.Count.Log <dbl> 2.772589, 5.697093, 5.442418, 3.044522, 2.30…
## $ Hfr.or.Cow <dbl> 1, 0, 0, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 1, 0,…
## $ BS.Month_warm <dbl> 0, 1, 1, 0, 0, 0, 1, 0, 1, 1, 0, 1, 1, 0, 0,…
## $ CE.Stom.Tension01 <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
## $ CE.Stom.Fullness01 <dbl> 0, 1, 1, 1, 1, 0, 0, 0, 0, 1, 1, 1, 0, 0, 1,…
## $ CE.Lame <dbl> 1, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0,…
## $ BS.NEFA.0.7 <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0,…
## Observations: 712
## Variables: 10
## $ MS.Milk.Yield <dbl> 1.040073139, 1.387694150, -0.002789896, 0.95…
## $ MS.Protein <dbl> -1.55816442, -0.73740583, -1.36875859, 0.430…
## $ MS.Lactose <dbl> -0.50740817, 0.16796859, 0.73944124, -0.0398…
## $ MS.S.Cell.Count.Log <dbl> -1.08085884, 1.16364429, 0.96818536, -0.8721…
## $ Hfr.or.Cow <dbl> 0.6397853, -1.5608290, -1.5608290, 0.6397853…
## $ BS.Month_warm <dbl> -0.8800408, 1.1347150, 1.1347150, -0.8800408…
## $ CE.Stom.Tension01 <dbl> -0.2746318, -0.2746318, -0.2746318, -0.27463…
## $ CE.Stom.Fullness01 <dbl> -1.1444851, 0.8725282, 0.8725282, 0.8725282,…
## $ CE.Lame <dbl> 2.7509398, -0.3630016, -0.3630016, -0.363001…
## $ BS.NEFA.0.7 <dbl> -0.4943799, -0.4943799, -0.4943799, -0.49437…PCA and k-means clustering of hp11 check for missing values first, then perform PCA, plot and summarize Note: MS.Milk.Yield will be used instead of Milk.Yield (there are 325 missing values in Milk.Yield)
## MS.Milk.Yield MS.Protein MS.Lactose
## N 7.120000e+02 7.120000e+02 7.120000e+02
## mean -1.216243e-16 -4.765935e-16 5.121147e-16
## Std.Dev. 1.000000e+00 1.000000e+00 1.000000e+00
## min -2.403030e+00 -2.726167e+00 -1.032635e+01
## Q1 -6.483718e-01 -7.058382e-01 -5.074082e-01
## median -2.014305e-01 -7.448544e-02 6.406448e-02
## Q3 4.441514e-01 6.200026e-01 6.355371e-01
## max 5.194972e+00 8.511912e+00 2.298003e+00
## missing values 0.000000e+00 0.000000e+00 0.000000e+00
## MS.S.Cell.Count.Log Hfr.or.Cow BS.Month_warm
## N 7.120000e+02 7.120000e+02 7.120000e+02
## mean -1.845848e-16 2.337293e-17 -3.695820e-17
## Std.Dev. 1.000000e+00 1.000000e+00 1.000000e+00
## min -1.441578e+00 -1.560829e+00 -8.800408e-01
## Q1 -7.696719e-01 -1.560829e+00 -8.800408e-01
## median -2.063642e-01 6.397853e-01 -8.800408e-01
## Q3 6.238880e-01 6.397853e-01 1.134715e+00
## max 3.500504e+00 6.397853e-01 1.134715e+00
## missing values 0.000000e+00 0.000000e+00 0.000000e+00
## CE.Stom.Tension01 CE.Stom.Fullness01 CE.Lame
## N 7.120000e+02 7.120000e+02 7.120000e+02
## mean -1.972882e-17 -1.551579e-16 1.830139e-17
## Std.Dev. 1.000000e+00 1.000000e+00 1.000000e+00
## min -2.746318e-01 -1.144485e+00 -3.630016e-01
## Q1 -2.746318e-01 -1.144485e+00 -3.630016e-01
## median -2.746318e-01 8.725282e-01 -3.630016e-01
## Q3 -2.746318e-01 8.725282e-01 -3.630016e-01
## max 3.636125e+00 8.725282e-01 2.750940e+00
## missing values 0.000000e+00 0.000000e+00 0.000000e+00
## BS.NEFA.0.7
## N 7.120000e+02
## mean -5.049915e-17
## Std.Dev. 1.000000e+00
## min -4.943799e-01
## Q1 -4.943799e-01
## median -4.943799e-01
## Q3 -4.943799e-01
## max 2.019895e+00
## missing values 0.000000e+00
## Importance of components:
## Comp.1 Comp.2 Comp.3 Comp.4 Comp.5
## Standard deviation 1.2744174 1.2209574 1.1242494 1.0664119 0.99498377
## Proportion of Variance 0.1626424 0.1492834 0.1265714 0.1138834 0.09913851
## Cumulative Proportion 0.1626424 0.3119258 0.4384972 0.5523806 0.65151910
## Comp.6 Comp.7 Comp.8 Comp.9
## Standard deviation 0.92216926 0.87169483 0.84211875 0.78936246
## Proportion of Variance 0.08515922 0.07609206 0.07101614 0.06239695
## Cumulative Proportion 0.73667832 0.81277038 0.88378652 0.94618347
## Comp.10
## Standard deviation 0.73308216
## Proportion of Variance 0.05381653
## Cumulative Proportion 1.00000000Note: need 3 PC’s to reach 50% variance…not a strong clustering, but acceptable Note 2: Hapto.Log was NOT used for the clustering
these graphs are not very useful, but they show the variance explained by the clustering per PC 
## [1] 712 10save first 2 PC’s as data frame
## 2 groups 3 groups 4 groups 5 groups 6 groups 7 groups
## SSE 1498.602345 1025.879128 833.59207 697.02354 576.1348054 475.3007574
## ssi 1.075525 1.918493 1.64959 2.11894 0.6588912 0.8442455
## 8 groups 9 groups 10 groups 11 groups 12 groups 13 groups
## SSE 418.3641339 371.524826 330.331521 299.565731 272.5845031 251.0064518
## ssi 0.8779883 1.020716 1.073037 1.051912 0.9987275 0.9987275
## 14 groups 15 groups 16 groups 17 groups 18 groups 19 groups
## SSE 229.712189 212.210008 197.540219 182.555307 172.835935 164.376120
## ssi 1.418272 1.397646 1.414821 1.376215 1.357759 1.357759
## 20 groups
## SSE 155.802044
## ssi 1.662432
use 5 clusters based on maximum in ssi graph, could use 3 only the Silhouette plot is not working; and number of observations per cluster
##
## 1 2 3 4 5
## 173 94 195 82 168distribution of observations over 4 clusters is quite even
##
## 1 2 3 4 5
## 173 94 195 82 168
Note: cluster in the middle of the other clusters is linked to almost all other clusters; it is kind of a ‘fuzzy’ cluster
cbind cluster numbers into data set hp10, i.e. the non centered, non-scaled data set
cbind cluster numbers into data set hp9, the original data set, and save as RData set, as well
Plots and descriptives of variables per cluster
## 'data.frame': 712 obs. of 5 variables:
## $ MS.Milk.Yield : num 21.2 23.3 14.9 20.7 11.9 17 12.5 11.4 36.8 12.8 ...
## $ MS.Protein : num 2.79 3.05 2.85 3.42 2.89 2.76 3.3 3.7 2.86 3.37 ...
## $ MS.Lactose : num 4.73 4.86 4.97 4.82 4.81 4.8 4.83 4.81 4.82 4.85 ...
## $ MS.S.Cell.Count.Log: num 2.77 5.7 5.44 3.04 2.3 ...
## $ clusterHapto : int 1 2 2 4 2 1 3 4 1 3 ...
cbind the Hapto.Log and Hapto.0.35 varaibles into the data set
## Observations: 712
## Variables: 13
## $ MS.Milk.Yield <dbl> 21.2, 23.3, 14.9, 20.7, 11.9, 17.0, 12.5, 11…
## $ MS.Protein <dbl> 2.79, 3.05, 2.85, 3.42, 2.89, 2.76, 3.30, 3.…
## $ MS.Lactose <dbl> 4.73, 4.86, 4.97, 4.82, 4.81, 4.80, 4.83, 4.…
## $ MS.S.Cell.Count.Log <dbl> 2.772589, 5.697093, 5.442418, 3.044522, 2.30…
## $ Hfr.or.Cow <dbl> 1, 0, 0, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 1, 0,…
## $ BS.Month_warm <dbl> 0, 1, 1, 0, 0, 0, 1, 0, 1, 1, 0, 1, 1, 0, 0,…
## $ CE.Stom.Tension01 <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
## $ CE.Stom.Fullness01 <dbl> 0, 1, 1, 1, 1, 0, 0, 0, 0, 1, 1, 1, 0, 0, 1,…
## $ CE.Lame <dbl> 1, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0,…
## $ BS.NEFA.0.7 <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0,…
## $ clusterHapto <int> 1, 2, 2, 4, 2, 1, 3, 4, 1, 3, 3, 3, 2, 4, 4,…
## $ V2 <dbl> 6.392754, 5.634790, 6.608675, 5.634790, 7.55…
## $ V3 <fct> 0, 0, 0, 0, 0, 1, 1, 0, 1, 1, 0, 1, 1, 0, 0,…## 'data.frame': 712 obs. of 6 variables:
## $ MS.Milk.Yield : num 21.2 23.3 14.9 20.7 11.9 17 12.5 11.4 36.8 12.8 ...
## $ MS.Protein : num 2.79 3.05 2.85 3.42 2.89 2.76 3.3 3.7 2.86 3.37 ...
## $ MS.Lactose : num 4.73 4.86 4.97 4.82 4.81 4.8 4.83 4.81 4.82 4.85 ...
## $ MS.S.Cell.Count.Log: num 2.77 5.7 5.44 3.04 2.3 ...
## $ clusterHapto : int 1 2 2 4 2 1 3 4 1 3 ...
## $ Hapto.Log : num 6.39 5.63 6.61 5.63 7.55 ...
Note: could add any other numeric variable of interest to the graph
descriptives by cluster for MS.Milk.Yield
## $`1`
## [,1]
## N 94.000000
## mean 16.680851
## Std.Dev. 7.509224
## min 2.100000
## Q1 11.600000
## median 15.450000
## Q3 20.225000
## max 46.300000
## missing values 0.000000
##
## $`2`
## [,1]
## N 168.000000
## mean 14.112500
## Std.Dev. 4.841329
## min 0.600000
## Q1 11.000000
## median 13.300000
## Q3 15.700000
## max 40.600000
## missing values 0.000000
##
## $`3`
## [,1]
## N 173.000000
## mean 16.193642
## Std.Dev. 6.613799
## min 4.200000
## Q1 11.700000
## median 14.900000
## Q3 18.900000
## max 40.900000
## missing values 0.000000
##
## $`4`
## [,1]
## N 195.00000
## mean 12.99692
## Std.Dev. 4.59189
## min 0.40000
## Q1 10.10000
## median 12.20000
## Q3 15.10000
## max 32.70000
## missing values 0.00000
##
## $`5`
## [,1]
## N 82.000000
## mean 16.414634
## Std.Dev. 6.656439
## min 1.400000
## Q1 12.500000
## median 15.300000
## Q3 18.700000
## max 36.200000
## missing values 0.000000descriptives by cluster for MS.Protein
## $`1`
## [,1]
## N 94.0000000
## mean 3.0201064
## Std.Dev. 0.2232482
## min 2.4200000
## Q1 2.8700000
## median 3.0100000
## Q3 3.1600000
## max 3.6200000
## missing values 0.0000000
##
## $`2`
## [,1]
## N 168.0000000
## mean 3.1285119
## Std.Dev. 0.2256292
## min 2.6200000
## Q1 2.9575000
## median 3.1400000
## Q3 3.2725000
## max 3.6800000
## missing values 0.0000000
##
## $`3`
## [,1]
## N 173.0000000
## mean 3.2245665
## Std.Dev. 0.2253459
## min 2.7400000
## Q1 3.0600000
## median 3.2100000
## Q3 3.3700000
## max 3.9200000
## missing values 0.0000000
##
## $`4`
## [,1]
## N 195.0000000
## mean 3.5203077
## Std.Dev. 0.2495251
## min 2.9300000
## Q1 3.3700000
## median 3.5300000
## Q3 3.6800000
## max 4.2100000
## missing values 0.0000000
##
## $`5`
## [,1]
## N 82.0000000
## mean 3.4650000
## Std.Dev. 0.3944718
## min 2.8100000
## Q1 3.2750000
## median 3.4600000
## Q3 3.6175000
## max 5.9800000
## missing values 0.0000000descriptives by cluster for MS.Lactose
## $`1`
## [,1]
## N 94.0000000
## mean 4.8218085
## Std.Dev. 0.1557705
## min 4.2400000
## Q1 4.7325000
## median 4.8300000
## Q3 4.9200000
## max 5.1500000
## missing values 0.0000000
##
## $`2`
## [,1]
## N 168.000000
## mean 4.975893
## Std.Dev. 0.124676
## min 4.660000
## Q1 4.877500
## median 4.970000
## Q3 5.070000
## max 5.270000
## missing values 0.000000
##
## $`3`
## [,1]
## N 173.000000
## mean 4.846243
## Std.Dev. 0.116276
## min 4.520000
## Q1 4.770000
## median 4.840000
## Q3 4.920000
## max 5.100000
## missing values 0.000000
##
## $`4`
## [,1]
## N 195.0000000
## mean 4.7905641
## Std.Dev. 0.1500934
## min 4.2500000
## Q1 4.7050000
## median 4.8100000
## Q3 4.8900000
## max 5.1300000
## missing values 0.0000000
##
## $`5`
## [,1]
## N 82.0000000
## mean 4.5797561
## Std.Dev. 0.2660269
## min 2.8400000
## Q1 4.5125000
## median 4.6500000
## Q3 4.7100000
## max 4.9600000
## missing values 0.0000000descriptives by cluster for MS.S.Cell.Count
## $`1`
## [,1]
## N 94.000000
## mean 4.118352
## Std.Dev. 1.190854
## min 2.302585
## Q1 3.146134
## median 3.794921
## Q3 5.213506
## max 7.602401
## missing values 0.000000
##
## $`2`
## [,1]
## N 168.0000000
## mean 3.5054926
## Std.Dev. 0.9411548
## min 2.3025851
## Q1 2.8180572
## median 3.3140207
## Q3 4.0253517
## max 6.9641356
## missing values 0.0000000
##
## $`3`
## [,1]
## N 173.000000
## mean 4.063414
## Std.Dev. 1.059461
## min 2.302585
## Q1 3.258097
## median 3.970292
## Q3 4.634729
## max 7.540090
## missing values 0.000000
##
## $`4`
## [,1]
## N 195.000000
## mean 4.246548
## Std.Dev. 1.279004
## min 2.302585
## Q1 3.258097
## median 4.007333
## Q3 5.071803
## max 8.119696
## missing values 0.000000
##
## $`5`
## [,1]
## N 82.000000
## mean 5.728184
## Std.Dev. 1.316079
## min 2.890372
## Q1 4.884269
## median 5.798721
## Q3 6.354653
## max 8.741935
## missing values 0.000000descriptives by cluster for Hapto.Log
## $`1`
## [,1]
## N 94.000000
## mean 9.208024
## Std.Dev. 3.274849
## min 5.010635
## Q1 5.972134
## median 8.558402
## Q3 12.671557
## max 14.756045
## missing values 0.000000
##
## $`2`
## [,1]
## N 168.000000
## mean 6.728636
## Std.Dev. 1.788273
## min 4.976734
## Q1 5.634790
## median 5.938851
## Q3 7.415679
## max 13.774741
## missing values 0.000000
##
## $`3`
## [,1]
## N 173.000000
## mean 7.202421
## Std.Dev. 2.321140
## min 4.962845
## Q1 5.634790
## median 6.061457
## Q3 8.055951
## max 14.406542
## missing values 0.000000
##
## $`4`
## [,1]
## N 195.000000
## mean 6.739156
## Std.Dev. 2.137796
## min 4.948760
## Q1 5.634790
## median 5.634790
## Q3 6.830103
## max 13.778172
## missing values 0.000000
##
## $`5`
## [,1]
## N 82.000000
## mean 8.423959
## Std.Dev. 2.902437
## min 4.941642
## Q1 5.758496
## median 7.392439
## Q3 10.751709
## max 14.510207
## missing values 0.000000repeat graphs and descriptive stats for the factor variables
## 'data.frame': 712 obs. of 7 variables:
## $ Hfr.or.Cow : num 1 0 0 1 1 1 1 1 1 1 ...
## $ BS.Month_warm : num 0 1 1 0 0 0 1 0 1 1 ...
## $ CE.Stom.Tension01 : num 0 0 0 0 0 0 0 0 0 0 ...
## $ CE.Stom.Fullness01: num 0 1 1 1 1 0 0 0 0 1 ...
## $ CE.Lame : num 1 0 0 0 0 1 0 0 1 0 ...
## $ BS.NEFA.0.7 : num 0 0 0 0 0 0 0 0 0 1 ...
## $ clusterHapto : int 1 2 2 4 2 1 3 4 1 3 ...
## , , = Hfr.or.Cow
##
##
## 1 2 3 4 5
## 0 3.90 10.52 2.70 4.02 0.81
## 1 12.77 6.15 13.97 12.65 15.85
##
## , , = BS.Month_warm
##
##
## 1 2 3 4 5
## 0 3.90 6.75 9.25 14.44 9.35
## 1 12.77 9.92 7.42 2.22 7.32
##
## , , = CE.Stom.Tension01
##
##
## 1 2 3 4 5
## 0 16.13 16.17 15.80 15.04 13.82
## 1 0.53 0.50 0.87 1.62 2.85
##
## , , = CE.Stom.Fullness01
##
##
## 1 2 3 4 5
## 0 14.18 1.39 10.40 3.33 13.62
## 1 2.48 15.28 6.26 13.33 3.05
##
## , , = CE.Lame
##
##
## 1 2 3 4 5
## 0 8.33 16.57 14.84 16.58 13.62
## 1 8.33 0.10 1.83 0.09 3.05
##
## , , = BS.NEFA.0.7
##
##
## 1 2 3 4 5
## 0 5.50 15.08 12.81 16.67 12.40
## 1 11.17 1.59 3.85 0.00 4.27## 'data.frame': 60 obs. of 4 variables:
## $ levels : Factor w/ 2 levels "0","1": 1 2 1 2 1 2 1 2 1 2 ...
## $ clusterHapto: Factor w/ 5 levels "1","2","3","4",..: 1 1 2 2 3 3 4 4 5 5 ...
## $ variable : Factor w/ 6 levels "Hfr.or.Cow","BS.Month_warm",..: 1 1 1 1 1 1 1 1 1 1 ...
## $ proportion : num 3.9 12.77 10.52 6.15 2.7 ...
## , , = Hfr.or.Cow
##
##
## 1 2 3 4 5
## 0 22 106 28 47 4
## 1 72 62 145 148 78
##
## , , = BS.Month_warm
##
##
## 1 2 3 4 5
## 0 22 68 96 169 46
## 1 72 100 77 26 36
##
## , , = CE.Stom.Tension01
##
##
## 1 2 3 4 5
## 0 91 163 164 176 68
## 1 3 5 9 19 14
##
## , , = CE.Stom.Fullness01
##
##
## 1 2 3 4 5
## 0 80 14 108 39 67
## 1 14 154 65 156 15
##
## , , = CE.Lame
##
##
## 1 2 3 4 5
## 0 47 167 154 194 67
## 1 47 1 19 1 15
##
## , , = BS.NEFA.0.7
##
##
## 1 2 3 4 5
## 0 31 152 133 195 61
## 1 63 16 40 0 21Can add any other factor variable to this graph, need good notes to recall levels!!! 1 = event of interest
…………………………………………………… REPEAT clustering including Hapto.Log collect all ssign variables into hp10 and format factor variables as numeric dummy variables; check pair-wise correlations again
## Observations: 712
## Variables: 10
## $ MS.Milk.Yield <dbl> 21.2, 23.3, 14.9, 20.7, 11.9, 17.0, 12.5, 11…
## $ MS.Protein <dbl> 2.79, 3.05, 2.85, 3.42, 2.89, 2.76, 3.30, 3.…
## $ MS.Lactose <dbl> 4.73, 4.86, 4.97, 4.82, 4.81, 4.80, 4.83, 4.…
## $ MS.S.Cell.Count.Log <dbl> 2.772589, 5.697093, 5.442418, 3.044522, 2.30…
## $ Hfr.or.Cow <dbl> 1, 0, 0, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 1, 0,…
## $ BS.Month_warm <dbl> 0, 1, 1, 0, 0, 0, 1, 0, 1, 1, 0, 1, 1, 0, 0,…
## $ CE.Stom.Tension01 <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
## $ CE.Stom.Fullness01 <dbl> 0, 1, 1, 1, 1, 0, 0, 0, 0, 1, 1, 1, 0, 0, 1,…
## $ CE.Lame <dbl> 1, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0,…
## $ BS.NEFA.0.7 <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0,…
## Observations: 712
## Variables: 11
## $ MS.Milk.Yield <dbl> 1.040073139, 1.387694150, -0.002789896, 0.95…
## $ MS.Protein <dbl> -1.55816442, -0.73740583, -1.36875859, 0.430…
## $ MS.Lactose <dbl> -0.50740817, 0.16796859, 0.73944124, -0.0398…
## $ MS.S.Cell.Count.Log <dbl> -1.08085884, 1.16364429, 0.96818536, -0.8721…
## $ Hfr.or.Cow <dbl> 0.6397853, -1.5608290, -1.5608290, 0.6397853…
## $ BS.Month_warm <dbl> -0.8800408, 1.1347150, 1.1347150, -0.8800408…
## $ CE.Stom.Tension01 <dbl> -0.2746318, -0.2746318, -0.2746318, -0.27463…
## $ CE.Stom.Fullness01 <dbl> -1.1444851, 0.8725282, 0.8725282, 0.8725282,…
## $ CE.Lame <dbl> 2.7509398, -0.3630016, -0.3630016, -0.363001…
## $ BS.NEFA.0.7 <dbl> -0.4943799, -0.4943799, -0.4943799, -0.49437…
## $ Hapto.Log <dbl> -0.38442550, -0.68282945, -0.29941937, -0.68…PCA and k-means clustering of hp11 check for missing values first, then perform PCA, plot and summarize Note: MS.Milk.Yield will be used instead of Milk.Yield (there are 325 missing values in Milk.Yield)
## MS.Milk.Yield MS.Protein MS.Lactose
## N 7.120000e+02 7.120000e+02 7.120000e+02
## mean -1.216243e-16 -4.765935e-16 5.121147e-16
## Std.Dev. 1.000000e+00 1.000000e+00 1.000000e+00
## min -2.403030e+00 -2.726167e+00 -1.032635e+01
## Q1 -6.483718e-01 -7.058382e-01 -5.074082e-01
## median -2.014305e-01 -7.448544e-02 6.406448e-02
## Q3 4.441514e-01 6.200026e-01 6.355371e-01
## max 5.194972e+00 8.511912e+00 2.298003e+00
## missing values 0.000000e+00 0.000000e+00 0.000000e+00
## MS.S.Cell.Count.Log Hfr.or.Cow BS.Month_warm
## N 7.120000e+02 7.120000e+02 7.120000e+02
## mean -1.845848e-16 2.337293e-17 -3.695820e-17
## Std.Dev. 1.000000e+00 1.000000e+00 1.000000e+00
## min -1.441578e+00 -1.560829e+00 -8.800408e-01
## Q1 -7.696719e-01 -1.560829e+00 -8.800408e-01
## median -2.063642e-01 6.397853e-01 -8.800408e-01
## Q3 6.238880e-01 6.397853e-01 1.134715e+00
## max 3.500504e+00 6.397853e-01 1.134715e+00
## missing values 0.000000e+00 0.000000e+00 0.000000e+00
## CE.Stom.Tension01 CE.Stom.Fullness01 CE.Lame
## N 7.120000e+02 7.120000e+02 7.120000e+02
## mean -1.972882e-17 -1.551579e-16 1.830139e-17
## Std.Dev. 1.000000e+00 1.000000e+00 1.000000e+00
## min -2.746318e-01 -1.144485e+00 -3.630016e-01
## Q1 -2.746318e-01 -1.144485e+00 -3.630016e-01
## median -2.746318e-01 8.725282e-01 -3.630016e-01
## Q3 -2.746318e-01 8.725282e-01 -3.630016e-01
## max 3.636125e+00 8.725282e-01 2.750940e+00
## missing values 0.000000e+00 0.000000e+00 0.000000e+00
## BS.NEFA.0.7 Hapto.Log
## N 7.120000e+02 7.120000e+02
## mean -5.049915e-17 1.601124e-17
## Std.Dev. 1.000000e+00 1.000000e+00
## min -4.943799e-01 -9.557153e-01
## Q1 -4.943799e-01 -6.828294e-01
## median -4.943799e-01 -4.854276e-01
## Q3 -4.943799e-01 3.495475e-01
## max 2.019895e+00 2.908128e+00
## missing values 0.000000e+00 0.000000e+00
## Importance of components:
## Comp.1 Comp.2 Comp.3 Comp.4 Comp.5
## Standard deviation 1.3547281 1.2714653 1.1659407 1.0664121 1.00968509
## Proportion of Variance 0.1670791 0.1471725 0.1237572 0.1035304 0.09280889
## Cumulative Proportion 0.1670791 0.3142516 0.4380088 0.5415392 0.63434810
## Comp.6 Comp.7 Comp.8 Comp.9
## Standard deviation 0.92236077 0.87441102 0.84447153 0.78945510
## Proportion of Variance 0.07744963 0.06960636 0.06492138 0.05673781
## Cumulative Proportion 0.71179773 0.78140410 0.84632547 0.90306329
## Comp.10 Comp.11
## Standard deviation 0.74291139 0.71616262
## Proportion of Variance 0.05024487 0.04669184
## Cumulative Proportion 0.95330816 1.00000000Note: need 3 PC’s to reach 50% variance…not a strong clustering, but acceptable Note 2: Hapto.Log was NOT used for the clustering
these graphs are not very useful, but they show the variance explained by the clustering per PC 
save first 2 PC’s as data frame
## 2 groups 3 groups 4 groups 5 groups 6 groups 7 groups
## SSE 1553.5926985 1084.317669 856.218233 725.430869 606.8355549 498.7290525
## ssi 0.7589306 1.914652 1.206969 1.672581 0.5223833 0.6090285
## 8 groups 9 groups 10 groups 11 groups 12 groups
## SSE 420.8754657 377.4103279 339.3894363 305.8605196 277.9890049
## ssi 0.5992915 0.6208762 0.6175573 0.6631954 0.6666952
## 13 groups 14 groups 15 groups 16 groups 17 groups 18 groups
## SSE 254.3820623 233.5398266 217.3774045 203.260330 191.2415393 180.7887876
## ssi 0.7676859 0.7652891 0.9378094 0.715831 0.7740985 0.9852646
## 19 groups 20 groups
## SSE 171.1566582 161.99590
## ssi 0.9947309 1.00861
use 3 clusters based on maximum in ssi graph, could use 5 as well like before the Silhouette plot is not working; and number of observations per cluster
##
## 1 2 3
## 143 283 286distribution of observations over 4 clusters is quite even
##
## 1 2 3
## 143 283 286## Warning: Removed 1 rows containing missing values (geom_point).
Note: Hapto.Log, CE.Lame, BS.NEFA.0.7, MS.S.Cell.Count and BS.Month_warm are in the same direction and define one cluster on the right. The CE.Stom.Tension01, CE.Stom.Fullness01 and other variable axis are in the oposite direction defining the other two clusters
…. cbind cluster numbers into data set hp10, i.e. the non centered, non-scaled data set
cbind cluster numbers into data set hp9, the original data set, and save as RData set, as well
Plots and descriptives of variables per cluster
## 'data.frame': 712 obs. of 6 variables:
## $ MS.Milk.Yield : num 21.2 23.3 14.9 20.7 11.9 17 12.5 11.4 36.8 12.8 ...
## $ MS.Protein : num 2.79 3.05 2.85 3.42 2.89 2.76 3.3 3.7 2.86 3.37 ...
## $ MS.Lactose : num 4.73 4.86 4.97 4.82 4.81 4.8 4.83 4.81 4.82 4.85 ...
## $ MS.S.Cell.Count.Log: num 2.77 5.7 5.44 3.04 2.3 ...
## $ Hapto.Log : num 6.39 5.63 6.61 5.63 7.55 ...
## $ clusterHapto : int 2 2 2 3 2 1 2 3 1 2 ...
cbind the Hapto.Log and Hapto.0.35 varaibles into the data set
## Observations: 712
## Variables: 13
## $ MS.Milk.Yield <dbl> 21.2, 23.3, 14.9, 20.7, 11.9, 17.0, 12.5, 11…
## $ MS.Protein <dbl> 2.79, 3.05, 2.85, 3.42, 2.89, 2.76, 3.30, 3.…
## $ MS.Lactose <dbl> 4.73, 4.86, 4.97, 4.82, 4.81, 4.80, 4.83, 4.…
## $ MS.S.Cell.Count.Log <dbl> 2.772589, 5.697093, 5.442418, 3.044522, 2.30…
## $ Hfr.or.Cow <dbl> 1, 0, 0, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 1, 0,…
## $ BS.Month_warm <dbl> 0, 1, 1, 0, 0, 0, 1, 0, 1, 1, 0, 1, 1, 0, 0,…
## $ CE.Stom.Tension01 <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
## $ CE.Stom.Fullness01 <dbl> 0, 1, 1, 1, 1, 0, 0, 0, 0, 1, 1, 1, 0, 0, 1,…
## $ CE.Lame <dbl> 1, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0,…
## $ BS.NEFA.0.7 <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0,…
## $ Hapto.Log <dbl> 6.392754, 5.634790, 6.608675, 5.634790, 7.55…
## $ clusterHapto <int> 2, 2, 2, 3, 2, 1, 2, 3, 1, 2, 2, 2, 2, 3, 3,…
## $ V2 <fct> 0, 0, 0, 0, 0, 1, 1, 0, 1, 1, 0, 1, 1, 0, 0,…## 'data.frame': 712 obs. of 6 variables:
## $ MS.Milk.Yield : num 21.2 23.3 14.9 20.7 11.9 17 12.5 11.4 36.8 12.8 ...
## $ MS.Protein : num 2.79 3.05 2.85 3.42 2.89 2.76 3.3 3.7 2.86 3.37 ...
## $ MS.Lactose : num 4.73 4.86 4.97 4.82 4.81 4.8 4.83 4.81 4.82 4.85 ...
## $ MS.S.Cell.Count.Log: num 2.77 5.7 5.44 3.04 2.3 ...
## $ Hapto.Log : num 6.39 5.63 6.61 5.63 7.55 ...
## $ clusterHapto : int 2 2 2 3 2 1 2 3 1 2 ... Note: could add any other numeric variable of interest to the graph
descriptives by cluster for MS.Milk.Yield
## $`1`
## [,1]
## N 143.000000
## mean 13.697203
## Std.Dev. 5.425695
## min 0.400000
## Q1 10.500000
## median 12.900000
## Q3 16.650000
## max 37.600000
## missing values 0.000000
##
## $`2`
## [,1]
## N 283.000000
## mean 14.879859
## Std.Dev. 5.416116
## min 0.600000
## Q1 11.100000
## median 13.600000
## Q3 17.150000
## max 40.600000
## missing values 0.000000
##
## $`3`
## [,1]
## N 286.000000
## mean 15.563287
## Std.Dev. 6.795877
## min 1.400000
## Q1 11.125000
## median 14.500000
## Q3 17.900000
## max 46.300000
## missing values 0.000000descriptives by cluster for MS.Protein
## $`1`
## [,1]
## N 143.0000000
## mean 3.1765035
## Std.Dev. 0.2998597
## min 2.4200000
## Q1 2.9750000
## median 3.1600000
## Q3 3.3900000
## max 3.9600000
## missing values 0.0000000
##
## $`2`
## [,1]
## N 283.0000000
## mean 3.1444523
## Std.Dev. 0.2200241
## min 2.6200000
## Q1 2.9700000
## median 3.1400000
## Q3 3.2800000
## max 3.8900000
## missing values 0.0000000
##
## $`3`
## [,1]
## N 286.0000000
## mean 3.4748252
## Std.Dev. 0.3097874
## min 2.7400000
## Q1 3.2800000
## median 3.4700000
## Q3 3.6400000
## max 5.9800000
## missing values 0.0000000descriptives by cluster for MS.Lactose
## $`1`
## [,1]
## N 143.0000000
## mean 4.7400699
## Std.Dev. 0.1795201
## min 4.2400000
## Q1 4.6600000
## median 4.7700000
## Q3 4.8500000
## max 5.1000000
## missing values 0.0000000
##
## $`2`
## [,1]
## N 283.0000000
## mean 4.9433569
## Std.Dev. 0.1273086
## min 4.5200000
## Q1 4.8550000
## median 4.9300000
## Q3 5.0400000
## max 5.2700000
## missing values 0.0000000
##
## $`3`
## [,1]
## N 286.0000000
## mean 4.7569930
## Std.Dev. 0.1954088
## min 2.8400000
## Q1 4.6700000
## median 4.7900000
## Q3 4.8600000
## max 5.1300000
## missing values 0.0000000descriptives by cluster for MS.S.Cell.Count
## $`1`
## [,1]
## N 143.000000
## mean 5.111426
## Std.Dev. 1.459594
## min 2.302585
## Q1 3.890987
## median 5.247024
## Q3 6.055595
## max 8.741935
## missing values 0.000000
##
## $`2`
## [,1]
## N 283.0000000
## mean 3.5697581
## Std.Dev. 0.9276868
## min 2.3025851
## Q1 2.8332133
## median 3.4657359
## Q3 4.1023359
## max 6.9641356
## missing values 0.0000000
##
## $`3`
## [,1]
## N 286.000000
## mean 4.320388
## Std.Dev. 1.223714
## min 2.302585
## Q1 3.367296
## median 4.085941
## Q3 5.153292
## max 8.674368
## missing values 0.000000descriptives by cluster for Hapto.Log
## $`1`
## [,1]
## N 143.000000
## mean 10.883418
## Std.Dev. 2.612452
## min 5.634790
## Q1 8.891110
## median 11.596163
## Q3 13.026715
## max 14.756045
## missing values 0.000000
##
## $`2`
## [,1]
## N 283.000000
## mean 6.649202
## Std.Dev. 1.665600
## min 4.976734
## Q1 5.634790
## median 5.823046
## Q3 7.313049
## max 13.774741
## missing values 0.000000
##
## $`3`
## [,1]
## N 286.000000
## mean 6.324584
## Std.Dev. 1.464240
## min 4.941642
## Q1 5.634790
## median 5.634790
## Q3 6.503165
## max 12.508656
## missing values 0.000000Note: if we defined a new cut-off for Hapto.Log above the minimum of cluster 1, exp(5.63479) = 280 This is a very low value compared to the other cut-offs from literature…but…
We could separate cluster 1 from the other two clusters….cluster 1 is the one defined by the disease variables… Would this be the starting point for a useful prediction model?
#……….. # Repeat graphs for factor variables
## 'data.frame': 712 obs. of 7 variables:
## $ Hfr.or.Cow : num 1 0 0 1 1 1 1 1 1 1 ...
## $ BS.Month_warm : num 0 1 1 0 0 0 1 0 1 1 ...
## $ CE.Stom.Tension01 : num 0 0 0 0 0 0 0 0 0 0 ...
## $ CE.Stom.Fullness01: num 0 1 1 1 1 0 0 0 0 1 ...
## $ CE.Lame : num 1 0 0 0 0 1 0 0 1 0 ...
## $ BS.NEFA.0.7 : num 0 0 0 0 0 0 0 0 0 1 ...
## $ clusterHapto : int 2 2 2 3 2 1 2 3 1 2 ...
## , , = Hfr.or.Cow
##
##
## 1 2 3
## 0 4.80 6.41 1.60
## 01 0.00 0.00 0.00
## 02 0.00 0.00 0.00
## 1 9.49 7.87 12.69
##
## , , = BS.Month_warm
##
##
## 1 2 3
## 0 4.90 6.11 11.54
## 01 0.00 0.00 0.00
## 02 0.00 0.00 0.00
## 1 9.39 8.18 2.75
##
## , , = CE.Stom.Tension01
##
##
## 1 2 3
## 0 12.79 13.73 13.09
## 01 0.00 0.00 0.00
## 02 0.00 0.00 0.00
## 1 1.50 0.56 1.20
##
## , , = CE.Stom.Fullness01
##
##
## 1 2 3
## 0 10.89 3.74 6.24
## 01 0.00 0.00 0.00
## 02 0.00 0.00 0.00
## 1 3.40 10.55 8.04
##
## , , = CE.Lame
##
##
## 1 2 3
## 0 9.09 13.28 13.74
## 01 0.00 0.00 0.00
## 02 0.00 0.00 0.00
## 1 5.19 1.01 0.55
##
## , , = BS.NEFA.0.7
##
##
## 1 2 3
## 0 6.49 11.86 13.59
## 01 0.00 0.00 0.00
## 02 0.00 0.00 0.00
## 1 7.79 2.42 0.70
##
## , , = hp9$Cow.Breed
##
##
## 1 2 3
## 0 0.00 0.00 0.00
## 01 10.19 10.15 6.69
## 02 4.10 4.14 7.59
## 1 0.00 0.00 0.00## 'data.frame': 84 obs. of 4 variables:
## $ levels : Factor w/ 4 levels "0","01","02",..: 1 2 3 4 1 2 3 4 1 2 ...
## $ clusterHapto: Factor w/ 3 levels "1","2","3": 1 1 1 1 2 2 2 2 3 3 ...
## $ variable : Factor w/ 7 levels "Hfr.or.Cow","BS.Month_warm",..: 1 1 1 1 1 1 1 1 1 1 ...
## $ proportion : num 4.8 0 0 9.49 6.41 0 0 7.87 1.6 0 ...
## , , = Hfr.or.Cow
##
##
## 1 2 3
## 0 48 127 32
## 01 0 0 0
## 02 0 0 0
## 1 95 156 254
##
## , , = BS.Month_warm
##
##
## 1 2 3
## 0 49 121 231
## 01 0 0 0
## 02 0 0 0
## 1 94 162 55
##
## , , = CE.Stom.Tension01
##
##
## 1 2 3
## 0 128 272 262
## 01 0 0 0
## 02 0 0 0
## 1 15 11 24
##
## , , = CE.Stom.Fullness01
##
##
## 1 2 3
## 0 109 74 125
## 01 0 0 0
## 02 0 0 0
## 1 34 209 161
##
## , , = CE.Lame
##
##
## 1 2 3
## 0 91 263 275
## 01 0 0 0
## 02 0 0 0
## 1 52 20 11
##
## , , = BS.NEFA.0.7
##
##
## 1 2 3
## 0 65 235 272
## 01 0 0 0
## 02 0 0 0
## 1 78 48 14
##
## , , = hp9$Cow.Breed
##
##
## 1 2 3
## 0 0 0 0
## 01 102 201 134
## 02 41 82 152
## 1 0 0 0